2244 J . Org. Chem., Vol. 36, No. 16, 1971
SCHWEIZER AND GREASY
was obtained which consisted of about 50% 3-isopropyl-2-nitrothiophene and 50% 4-isopropyl-2-nitrothiophene. Anal. Calcd for C J ~ ~ N O Z SC,: 49.11; H, 5.30; N, 8.18; S, 18.72. Found: C, 49.48; H, 5.35; N, 8.20; S, 18.59. The isomers were separated by preparative glpc (DEGS, 6 ft, 170"): 2-nitro-3-isopropylthiophene, nmr (CC14) 8 1.28 (d, 6, J = 7.0 He, CHs), 3.98 (complex m, 1, methine H), 7.03 (d, 1, J = 5.5 He, aromatic H ) , and 7.38 (d, 1, J = 5 Hz, aromatic H); 2-nitro-4-isopropylthiophene, nmr (CCL) S 1.27 (d, 6, J = 7.5 Hz, CHs), 2.98 (complex m, 1, methine H), 7.19 [d (slightly split), 1, J = 2.0 Hz, 5 H), and 7.76 (d, 1, J = 2.0 Ha, 3 H), and uv (95% EtOH) 289 mp (log e 3.76) and 327 (3.80). 3-Cyclopropyl-2-phenylthiophene (22) was prepared by irradiating with a high pressure mercury lamp 21 (500 mg, 2 mmol) in 125 ml of benzene containing 1 mol % anhydrous NazSzOs. After 6 hr irradiation stopped. The benzene solution was washed repeatedly with aqueous NalSzOa and then dried over MgS04. After removal of the solvents the residue was chromatographed over A1208 using benzene to give chiefly 22 (50% yield from glpc) which was purified by preparative glpc (F & M 700, 150') to give a clear liquid: ir (neat) 1035, 1056 cm-l (cyclopropyl); nmr (CCl,) S 0.5-1.0 (complex multiplet, 4, cyclopropylmethylene), 1.7-2.2 (complex multiplet, 1, tert-H), 6.55 [d, 1, J = 5.5 HI, 4(?) HI, 7.04 [d, 1, J = 5.5 Hz, 5(?) HI, and 7.1-7.6 (complex multiplet, 5, phenyl). Too little sample was obtained for an elemental analysis. Attempted Synthesis of 9 from Quaternary Iodide 7.-In a manner analogous to that described for 1, condensation of 3thiophenealdehyde with malonic acid gave 3-(3-thienyl)acrylic acid.** Reduction with LiAlH4 gave 3-(3-thienyl)-l-propanolin 66% yield, bp 141-145' (12 mm). Treatment with tosyl chloride gave an oily tosylate which was not purified but allowed to react with dimethylamine to give 3(3-thienyl)-l-dimethylaminopropene in 32% yield based on alcohol: bp 104-107" (12 mm); n l * ~ 1.5130. Treatment with methyl iodide gave in 90% yield the quaternary iodide (7), mp 167-169'. Anal. Calcd for CloHl&NI: C, 38.59; H, 5.84; S, 10.30; N, 4.50; I, 40.78. Found: C, 38.45; H , 5.71; S, 10.30; N , 4.40; I, 40.73. Attempted cyclization in liquid ammonia with sodium amide gave, after work-up, aliquid (about 30%), bp 60" (10 mm), which,
besides some weak signals ultimately attributed to 9, showed strong absorptions a t 8 6.3-6.5 indicating vinyl protons presumed to arise from 3-(3-thienyl)-l-propene (8). The reaction mixture was not investigated further.
Registry No.-1, 26019-23-8; 2, 29481-20-7; 3, 29481-21-8; 4, 29481-22-9; 5, 29481-23-0; 6, 2948124-1; 7, 29481-25-2; 9, 29576-51-0; 10, 29481-26-3; 11, 29481-27-4; 12, 29481-28-5; 14, 29481-29-6; 15, 29576-52-1; 16, 29481-30-9; 17, 29481-31-0; 19, 29481-32-1; 20, 29481-33-2; 21, 29481-34-3; 22, 29481-35-4; 23, 29481-36-5; 24, 29481-37-6; 4-isopropylpyridine1 696-30-0; 3-(2-thienyl)-l-propanol, 19498-72-7; 3- (2- thienyl) - 1-propyl tosylate, 29488-39-9; 3-(2-thienyl)-n-propyldimethylamine, 23711-40-2; 2-isopropyl-3-nitrothiophene, 29488-413; 3-(5-phenyl-2-thienyl)-propenoic acid, 29488-424; 3-(5-phenyl-2-thienyl)-l-propanoltosylate, 2948843-5; dimethyl-n- [3-(5-phenyl-2-thienyl) Jpropylamine, 29488-44-6 ; 3-(5-cyclopropyl-2-thienyl)propenoicacid, 29488-45-7 ; dimethyl-n- [3-(5-cyclopropyl-2-thienyl) 1propylamine, 29488-46-8 ; 4-isopropyl-2-nitrothiophene, 29488-47-9; 3-(3-thienyl)-l-propanol, 20905-98-0; 3-(3thieny1)-1-dimethylaminopropane, 29488-48-0; 4-cyclopropylpyridine, 4904-21-6. Acknowledgment.-We are extremely indebted to Miss N. Gjgs and Professor S. Gronowitz of the University of Lund, Sweden, for providing samples of nitrated phenylthiophenes. Helpful discussions with Professors Hans Wynberg and M. J. Janssen of this department are gratefully acknowledged. A portion of this work was initiated at Hope College, Holland, Mich.
Reactions of Phosphorus Compounds. XXVI. Preparation and Reactions of 3- and 4- Substituted 5-Benzoyl-2,2,2,5-tetraphenyloxa-2-phospholanes' EDWARD E. SCHWEIZER" AND WALTER S. GREASY Department of Chemistry, University of Delaware, Newark, Delaware 19711 Received January 6, 1971 Substituted (1- and 2-) vinylphosphonium salts reacted with the sodium salt of benzoin in DMSO to give 3and 4-substituted 5-benzoyl-2,2,2,5-tetraphenyloxa-2-phospholanes,respectively. Methyl-substituted oxaphospholanes were separated into diastereoisomers while 3- and 4-phenyl-substituted oxaphospholanes were NOlated as single isomers. Diastereomeric mixtures of oxaphospholanes substituted a t Cs tended to isomerize to one stable isomer while those substituted a t C4 did not isomerize. Reaction of 1-and 2-substituted vinylphosphonium salts with the sodium salt of benzoin in acetonitrile gave 2,5-dihydrofurans in all cases except 1-phenylvinyltriphenylphosphonium bromide which gave only the oxaphospholane. Pyrolysis of the oxaphospholanes gave benail, triphenylphosphine, and an olefin.
I n a previous article2 we described the preparation and reactions of 5-benzoyl-2,2,2,5-tetraphenyloxa-2phospholane (5) formed by attack of the benzoin enolate carbanion l a on vinyltriphenylphosphonium bromide (2) in dimethyl sulfoxide (DMSO). I n contrast, the dihydrofuran 3 was exclusively formed when acetonitrile was used as solvent (Scheme I).2 We now wish to report the reactions of 1 with 1- and 2-substituted vinylphosphonium salts which yield oxaphospholanes and dihydrofurans. The basis for the (1) E. E. Sohweizer and W. S. Creasy,'J. Org. Chem., 86, 2379 (1971). (2) E. E.Sohweizer, W. S. Creasy, J. G. Liehr, M. E. Jenkins, and D. L. Dalrymple, kbid., 86, 601 (1970).
stereochemical assignments of the oxaphospholanes obtained is discussed. The fusion reaction products of the oxaphospholanes are also examined. Previous workers have had variable success in obtaining products from conjugate additions to propenyltriphenylphosphonium bromide8 6. When 1 was allowed to react with this salt in DMSO, a 40% yield of oxaphospholanes was realized, which was easily separated into two diastereomeric oxaphospholanes 7a,b by selective extraction with chloroform and fractional crys(3) (a) P. Keough and M. Grayson, ibid., 18, 631 (1964); ( b ) D. Sesferth and J. Fogel, J. Organometal. Chem., 6, 205 (1866).
J . Org. Chem., Vol. 56, No. 16, 1971 2245
REACTIONS OF PHOSPHORUS COMPOUNDS SCHEME I 0Ph\ 0 CH
1
C
+
+
CH,=CHP(Ph),Br2
Ph'
-
Ph
+ (Ph),PO
'KGN
Ph
4
3
lb
1
~
COH
Ph
-
'F-0
11
A
c
DMSO ph/)
.OH
Ph'
bCH2CH$(Ph),
-
Ph'
OH
la
0 R'
II
+
+ R'CH==CR2P(Ph)$Br6,R'=CH3;R2=H
Ph'
8,R1-Ph;R2=H H R2a CHS 12,Rl=H;RZ=Ph
10, R'
CH Ph'
II
5b
R2
+ Ph Ph
'OCHR'
+ 4
Ph
14,R' = CH3; R2 = H
15,R1 =Ph; R2=H 16, R' =H; R2= CH, 17, R' = H, R2 Ph
-
CH,
--+
(Ph)2CHOH
ingly, due to proximity of the phenyl group on C5. Although 7a was insufficiehtly soluble for determination of proton nmr fine structure, it may very well be the other diastereoisomer as indicated. The predominance of 7a is probably a reflection, then, of the pseudoaxial position of methyl on the oxaphospholane ring of 7b.
-
phy==JI
7a
-
C-0
--c
C ~ P P h b Ph
WPhh
/""
I 0-
Ph,
7a, b, R1= CH,; R2= H 9, R' -Ph; R2= H lla, b, R' = H R2 CH3 13, R' = H R2= Ph
ti c-0 I
CCH&H2P(Ph)3
5a
tallization. The structural identity of these two compounds was related by basic hydrolysis to identical mixtures of the same product^.^ Infrared spectra of 7a and 7b are identical but the methyl absorptions in the
1
--C + I
+
0
I II + H02CCHCH2P(Ph),
-C
Reaction of benzoin (1) and 2-phenylvinyltriphenylphosphonium bromide (8) gave only one diastereoisomer 9. The Cq methine proton of 9 after deuterium exchange had 6 4.6 ppm and J a C C p = 11 Hz, indicating a gauche relationship to phosphorus. Compound 9 then corresponds in structure to 7a. Phenyl would be expected to have steric requirements similar to methyl with respect to protons. Compounds 7a, 7b, and 9 were allowed to react with hydrobromic acid to give the corresponding hydroxyphosphonium bromides 19a, 19b, and 20, respectively.
7b
18
proton nmr are significantly different. That of 7b appears at 0.6 ppm, considerably shielded with respect to the absorption observed in 7a (1.1 ppm). Compound 7b undergoes a deuterium for proton exchange at C3 when a sample in CDCla is shaken with DzO (requiring 48 hr), a reaction similarly reported for 5.2 The remaining methine proton, on C,, then appears as a doublet of quartets at 6 3.4 ppm with a vicinal P-H coupling constant of 34 Hz. This agrees very well with data reported for HAin compound 5 , with the relatively large coupling constant indicating an anti relationship to phosphorus.6 Methyl should then occupy the same position as HB in compound 5 and is shielded, accord(4) A mechanism for this reaction has been proposed in ref 2. (6) (a) J. G. Verkade, R. W. King, and C. W. Heitsch, Inorg. Cham., 8, 886 (1964); (b) J. G. Verkade and R. TV. King, zbzd., 1, 948 (1962).
7a, b
.-
HBr NaOH
ph'b\
,&I
ph4y
'CH26( Ph),BrOH 19a, b
There were only very small changes in position of the nmr methyl absorptions when 7a and 7b mere converted to their salts. These salts, therefore, retain the basic geometry of the oxaphospholanes from which they are 0 9 -
HBr
NaOH
c
c
Ph ,.*H
1
+
ph' ph? %OH H,P(Ph),Br20
2246 J. Org. Chem., Col. $6, No. 16,1871
SCHWEIZER AND CREASY
derived. On basification good reconversion to the parent oxaphospholanes was observed with no destruction of the stereochemistry of the ring, Le., no conversion to the opposite isomer. Reaction of 1 and l-methylvinyltriphenylphosphonium bromide (10)afforded 70% of a 1 :1 mixture of two diastereomeric oxaphospholanes, 1 la,b, separated by fractional crystallization. Compound l la had methyl absorption a t 6 1.4 ppm (d d) and compound llb showed an identical pattern centered at 6 1.2 ppm. Both patterns collapsed to doublets on shaking the sample with DzO and absorptions due to methylene protons on Cq appear as shown in Table I.
treatment with alkali a mixture consisting of 20% of lla and 80% of llb (78% overall). The impurity must arise from selective protonation since llb showed no tendency to isomerize to lla.
llb
21a
HBr
_._*
TABLE I PROTON NMRSPECTRAOF 5-BENZOYL-3-METHYL2,2,2,5-TETRAPHENYLOXA-2-PIIOSPHOLANES
lla
7
Proton
a, ppm
HA HB
3.4 2.3
7 , -
Hz
35 9
OH
JHCCP,
Proton
HA HB~
a, ppm
H5
3.3 2.0
20 27
Oxaphospholane 1 la exhibits proton nmr patterns and coupling constants quite similar to 5. However, the magnitudes of P-H coupling constants for protons of llb differ from other oxaphospholanes described, although there is little difference in chemical shifts between protons of l la and l l b. A reasonable explanation is as follows. The methyl in lla is pseudoequatorial on the ring which is conformationally similar to 5. However, to accommodate the methyl group in an equatorial position, the ring of llb must undergo a conformational change which then places HA! essentially gauche and HE/ anti to phosphorus, resulting in larger P-H coupling to the higher field proton H B ~ . Notably, the sum of P-H coupling constants in llb (44 Hz) is identical with that in lla (44 Hz). 0
Ila, R = CH,, 13,R=Ph
A
--.
llb--
JHCCP,
llb. R = CH;I
Furthermore, lla was found to isomerize slowly to llb on standing in chloroform. Conversion was 50% after 24 hr and required more than 1 week to reach the limits of nmr sensitivity. The compounds are relatively stable in the solid state but have identical melting points and mixture melting point, suggesting rapid isomerization at elevated temperatures. The observance of interconversion further confirms earlier observations concerning the presence of the phosphonium ylide in equilibrium with the oxaphospholane,2 although 31P nmr spectra consistently indicate that the latter form is predominant a t 25' in chloroform. Reaction of lla with hydrobromic acid yielded one pure salt 21a, which was converted into pure lla by aqueous base (85% overall) with no apparent isomerization. Similar treatment of llb gave, however, a mixture of salts containing 21% of 21a, which gave on
A
lla (2w0)
CH, 21b
The oxaphospholane 13, formed from reaction of 1 and 1-phenylvinyltriphenylphosphonium bromide (12), is isolated as a single diastereoisomer. The methylene protons on C? exhibit coupling constants and patterns similar to 5, after deuterium exchange (rapid). The low-field proton (6 4.1 ppm) had J H ~ C C =P 38 Ha and the high-field proton (6 2.55 ppm) had J H ~ C C=P10 Hz. Compound 13 is, therefore, one isomer as opposed to a rapidly equilibrating mixture and corresponds in structure t o lla. Reaction of 13 with hydrobromic acid afforded an apparent 1 : 1 mixture of phosphonium salts 22a,b which could not be separated into two components but which was converted to one pure salt, 22a, on boiling in benzene or warming with a catalytic amount of potassium tert-butylate in DMSO (quantitative recovery). Basic treatment of either the mixture or pure 22a gave pure 13 (74 and 67%, respectively), indicating that 13 is probably the more stable of the two possible isomers formed by the reaction of 1 and 12, the other isomerizing rapidly upon formation. These results suggest the phosphonium ylide is the form of 13 which undergoes protonation and that this oxaphospholane possesses significantly more ylide character than other oxaphospholanes examined. Indeed, the 31P nmr absorption a t +43.4 ppm is the smallest positive shift reported here, consistent with the lowest pentavalent phosphorus character. The increased acidity of benzylphosphonium salts as compared t o ethyl and propyl salts6 is undoubtedly responsible for the observed behavior of 13.
bH 22a, b
Oxaphospholanes '7a,7b,9,lla, llb,and 13 thermally decomposed to benzil (23)) and triphenylphosphine (24) on a vpc column at temperatures between 150 and 250". I n addition, styrene was detected in the decom(6) A. W. Johnson, "Ylid Chemistry," Academic Press, New York, N. Y., 1966.
J . Org. Chem., Vol. 36, No. 16,197i 2247
REACTIONS OF PHOSPHORUS COMPOUNDS
give 44.7 g of mixed oxaphospholanes 7a and 7b. Extraction of the mixture with chloroform followed by filtration gave a residue of 7a, pure by nmr, and a solution containing 7b. The latter was recovered in pure form by crystallization in two fractions from chloroform-hexane. Yields were 3.4 g of 7a (34%), mp 206-206.5', and 5.8 g (6%)' of 7b, mp 174- 175', mmp (between 7a and 7b) 174-187'. 5-Benzoy1-4-methyl-2,Z,2,5-tetraphenyloxa-Z-phospholane (7a): ir (Nujol) v 1075 (s, COC), 1110 (m, CP), 1155 (m, POC), 1235 (s), 1655 cm-l (s, C=O); nmr (AsCld) 6 1.1 (d, 3, CHI, J = 7 Hz), 2.8-3.5 (m, 3, HACCHZP),7.2-8.0 ppm (m, 25, CsHs) (the compound was not sufficiently soluble in any f R'HCsCHRZ :P(Ph)s other solvent); 31Pnmr (AsCls) 6 -21.3 ppm. Ph 24 Anal. Calcd for CasHaiOsP: C, 81.69; H, 6.07; P, 6.02. Found: C, 81,41; H, 5.98; P, 5.87. 23 5-Benzoyl-4-methyl-2,2,2,5-tetraphenyloxa-2-phospholane (7b): ir (Nujol) Y 1070 (s, COC), 1100 (m, CP), 1130 (6, POC), Similar results were obtained by pyrolysis of 5 in a 1220 (s, POC), 1655 cm-1 (s, C=O); nmr (CDCl,) 60.6 (d, 3, hot tube pyrolysis apparatus,' The oxaphospholanes CHa, J = 7 Elz), 2.5-3.9 (m, 3, HACCHZP),6.7-7.7 -~ ppm were relatively stable a t temperatures of refluxing tert(m, 25, CeHs). After the samde was shaken with DzO and allowed to stand butyl alcohol. 4 8 h r , the follo&ng spectrum was observed: 6 0.6 (d, 3, CHI), The vinylphosphonium salts 6, 8, 10, and 12 were 3.4 (doublet of quartets, 1, HACCP, J H ~=P 34 Hz), 6.7-7.7 allowed t o react with benzoin (1) in acetonitrile in order pprn (m, 25 c+). to prepare corresponding 4- and 5-substituted 2,3-diWhen 7b was impure with 7a, a methyl doublet corresponding phenyl-2,5-dihydrofurans.* The results are shown in to that described for the latter could be seen a t high amplitude. 3lP nmr (AsCls) showed 6 -21.5 ppm, (CHCl,) 6 4-56.5 ppm. Table 11. Anal. Calcd for CssHaiOzP: C, 81.69; H, 6.07; P, 6.02. Found: C, 81.66; H, 6.06; P, 6.05. TABLE I1 From the DMSO reaction filtrate was isolated 5 g of impure 2,3-diphenyl-5-methyl-2,5-dihydrofuran(14) which oxidized on PREPARATION O F 2,5-DIHYDROFURANS work-up, as described in a later experiment, to 1,2-diphenyl-2Dihydrofuran pentene-1,4-dione, 4.8 g, corresponding to a 10% yield of the Salt (% yield) Oxaphospholanes (%) dihydrofuran. When allyltriphenylphosphonium bromide was 6 14 (30) 7a 7b (32) substituted for 6, identical results were obtained. 8 15 (50) 9 (8) Hydrolysis of 7a and 7b in Aqueous-Ethanolic Sodium Hy10 16 (28) lla l l b (34) droxide.-A sample of 7a, 1.0 g (0.0002 mol), was allowed to 12 17 (0) 13 (76) reflux 72 hr in a mixture of 15 ml of ethanol and 15 ml of 20% aqueous sodium hydroxide. The solution was diluted with 100 ml of water and exbracted with two 50-ml portions of chloroIt is surprising that 15 was formed in highest yield form which was dried (MgSO,) and concentrated to give benzsince the Michael addition would be expected to enhydrol, 0.28 g (82%). Acidification of the aqueous solution followed by extraction with chloroform, drying (MgSO,), and counter greatest hindrance in attack on 8. Isolation concentration afforded 2-carboxypropyldiphenylphosphine oxide of 13 only from reaction of 1 and 12 is probably a re(18): 0.43 g (770/,); mp 135-137', after recrystallization from flection of the decreased reactivity of the benzyl ylide acetonitrile-ether (lit.$ 136'); ir (CHCla) Y 1120 (s, CP), 1160 (counterpart of sa) and greater hindrance to attack on (s, P=O), 1720 cm-l (s, C=O); nmr (CDCls) 6 1.3 (d, 3, CHI), 2.2-3.2 (m, 3, CHCHt), 7.2-8.0 (m, 10, CBH6),10.5 ppm the carbonyl, slowing intramolecular Wittig reaction (s, 1, COOH) and allowing effective competition for 12 by enolate la. Similar treatment of 7a, 1.0 g, gave benzhydrol, 0.26 g (77%), The latter observations are probably interrelated and and 2-carboxypropyldiphenylphosphine oxide (18),0.41 g (73%). provide valuable insight into the nature of the very 2-Phenylvinyltriphenylphosphonium Bromide (8).lo-lz-Tricomplex interacting equilibria. phenyl phosphine (24), 13.1 g (0.05 mol), and anhydrous aluminum chloride, 6.7 g (0.05 mol), were fused a t 150-200' for 2 hr and 2-bromostyrene, 13.8 g $0.078 mol), was added. The Experimental Section mixture was stirred at 180' for 4 hr and then a t 120" for 2 days. The remaining aluminum chloride complex was decomposed Infrared spectra were obtained on a Perkin-Elmer Infracord with water and organic products were extracted with chloroform 137 proton nmr spectra on a Varian A-60A analytical nmr which was dried (MgSO1), boiled with decolorizing charcoal, spectrometer using tetramethylsilane as standard and 3lP nmr filtered, and concentrated. Salt 8 was recovered by precipitaspectra on a Varian HR-60 nmr spectrometer using an 85y0 tion in anhydrous ether and was recrystallized several times phosphoric acid capillary as reference. Melting points are from acetone-ether. The yield, after a final recrystallization uncorrected and were obtained with a Thomas-Hoover capillary from chloroform-ether, was 12.1 g (54%), mp 263-265'." melting point apparatus. Analysis are by M-H-W Labora5-Benzoyl-2,2,2,4,5-pentaphenyloxa-2-phospholane ( 9 ).-To a tories, Garden City, Mich. Unless otherwise indicated all prereacted mixture of NaH, 0.9 g (0.02 mol), and 50 ml of dry reactions were undertaken in anhydrous conditions under a DMSO was added 1, 6.4 g (0.03 mol). After this stirred for blanket of dry nitrogen. Sodium hydride used was a 52% 15 min at ambient temperature, a solution of salt 8, 9.0 g (0.02 dispersion in mineral oil obtained from Metal Hydrides, Inc., mol, freshly dried at 100' for 24 hr), in 50 ml of DMSO was added Beverly, Mass. slowly and the reaction mixture allowed to stir 24 hr. No preS-Benzoyl-4-methy1-2,2,2,5-tetraphenyloxa-2-phospholanes cipitate formed but the green color of the sodium salt of benzoin (7a and 7b).-To a prereacted (1 hr at 40') mixture of NaH, faded slowly leaving a clear, tan solution which was diluted with 9.6 g (0.2 mol), and dry DMSO (200 ml) was added 4.4 g of 300 ml of water and extracted with two 150-ml portions of benzoin (1, 0.2 mol). When frothing had essentially ceased, chloroform. These were combined, dried over molecular salt 6, 77 g (0.2 mol), was introduced dropwise as a solution in sieves, filtered, concentrated to approximately 75 ml, and DMSO with the green color giving way near the equivalence dropped slowly into ether (anhydrous). The resulting precipipoint t o a light green precipitate. The reaction was stirred 6 hr at ambient temperature and filtered, and the residue washed (9) H.Hoffman, Chem. Ber., 94, 1331 (1961). with small amounts of DMSO, water, methanol, and ether to
position of 9 and 13. The apparent mechanism involves fragmentation of the five-membered ring into two T systems and a neutral phosphine.
+
"1;
+ +
( 7 ) Unpublished results, W. S. Creasy. (8) E. E. Sohweizer and J. G . Liehr, J. OTU.Chem., 88, 583 (1968).
(10) J. Chatt and F. G . Mann, J . Chem. SOC.,1192 (1940). (11) D.W.Allen and J. C. Tebby, Tetrahedron, 88, 2795 (1967). (12) S. Trippett and E. J. Walker, J. Chem. SOC.C , 887 (1966).
2248
J. Org. Chem., Vol. $6,No. 18, 1971
SCHWEIZER AND CREASY
tate w&s collected by filtration yielding 20, 1.6 g (12%), mp Anal. Calcd for Ca5HalOzP: C, 81.67; H, 6.07; P, 6.02. 176-179', identified by ir and nmr (described elsewhere). ConFound: C, 81.74; H , 5.91; P, 6.20. centration of the ether filtrate, dilution with an equal volume of 5-Benzoyl-3-methyl-2,2,5-tetraphenyloxa-2-phospholanes methanol, and cooling at 0' produced very slow crystallization ( l l a ) : ir (CHC13) was identical with that reported for I l b ; of oxaphospholane 9 which was dissolved in chloroform and nmr (CDCla) 6 1.4 (d d, 4, CHs), 2.1-2.6 (m, 1, HBCCP), recrystallized by the addition of hexane and cooling. The 2.7-3.9 (m, 2, HACCHP),6.9-8.0ppm (m, 25, CsHb). yield of 9 was 2.7 g (24%), mp 158-159'. This appeared When the above sample was shaken with DzO and allowed to to be one isomer only: ir (CHCla) v 1070 (s), 1085 (s, CO), stand 10 min, the following spectrum was observed: 6 1.4 (d, 1110 (s, CP), 1150 (m), 1180 (m, POC), 1240 (s), 1675 cm-l 3, CH3, J B ~ C C =P 21 Ha), 2.3 (d d , 1, HBCCP, J H ~=P 9 Hz), (s, C=O); nmr (CDCla) 6 3.1-3.9 (m, 2, CHZP),4.2-4.9 (m, 3.4 (d d, 1, HACCP, JH,\P = 35 He), 6.9-8.0 ppm (m, 25, 1, CHI, 6.6-7.8 ppm (m, 30, C&). c~H.5). When this sample was shaken with DzO and allowed to stand After 24 hr the sample had isomerized to a 50350 mixture of 48 hr, the following spectrum was observed: 6 4.6 (d, 1, HCCP, l l a and l l b , as determined by the nmr integration. After 10 JHP= 11 Ha), 6.6-7.8 ppm (m, 30, C6H5). 3lP nmr (CHCls) days the sample was essentially pure I l b . Solubility was not showed 6 f58.1 ppm. sufficient for determination of a nmr spectrum. Anal. Calcd for C40H330zP:C, 83.41; H, 5.77; P , 5.37. Anal. Calcd for C ~ ~ H ~ I O C, Z P 81.69; : H, 6.07; P , 6.02. Found: C, 83.19; H , 5.59; P, 5.74. Found: C, 81.72; H, 5.98; P, 5.88. The remaining solution was concentrated and the resulting 5-Benzoyl-2,2,2,3,5-pentaphenyloxa. 2-phospholane (13).-TO a prereacted mixture of NaH, 0.9 g (0.02 mol), and 50 ml of oil dissolved in a small amount of chloroform, diluted with hexane until turbid, and allowed to stand. White crystals of dry DMSO was added benzoin (11, 4.5 g (0.2 mol), and then (slowly) a solution of l2,l5 9.0 g (0.02 mol), in 50 ml of DMSO. 1,2-diphenylethyldiphenylphosphine oxide were deposited, mp A light green precipitate formed slowly and was collected by 228-230" (lit.I2 232-233'), 3.9 g (51%). This probably arises from basic hydrolysis of 8. filtration after 4 hr. The residue was washed with small amounts Concentration of the remaining solution gave an oil which of DMSO, water, acetonitiile, and ether and fractionally crysproved to be benzil (23) (probably via autoxidation of benzoin tallized from a chloroform-hexane mixture. All of these rein alkaline solution) contaminated with trace amounts of benzalcovered fractions, as well as the material recovered from washdehyde, styrene, and triphenylphosphine oxide, identified by ings of the initial residue, proved to be the same compound, 13: nmr, vpc, and tlc. The oil weighed 2.9 g, corresponding to combined yield 8.0 g (70%); mp 169-171'; ir (CHCla) v 1070 approximately 44y0 of the benzoin used. (s, CO), 1100 (m, CP), 1120 (m), 1170 (m, COP), 1230 (s), Total recovery of phosphorus was 87%. No evidence for 1660 cm-l (s, c=o); nmr (CL)Cla) 6 2.2-2.8 (m, 1 , HBCCP), the other oxaphospholane isomer was observed and, if present 3.5-4.6 (m, 2, HACCHP),6.8-8.0ppm (m, 30, C6H6). a t all, this must be formed in very small amounts. When this sample was shaken with DzO approximately 2 min, I-Methylvinyltriphenylphosphonium Bromide ( 10).1314-Arl. the following spectrum was observed: 6 2.5 (d d, 1, HBCCP, hydrous nickel bromide, 10.9 g (0.05 mol), and triphenylphosJHBP = 10, J H A H B = 14 He), 4.1 (d d, 1 , HACCP, J E ~ =P38 phine, 13.1 g (0.05 mol), were fused at 200' for 1 hr in a threeHz), 6.8-8.0 ppm (m, 30, C&). nmr (CHCla) showed 6 necked round-bottomed flask. Benzonitrile (100 ml) was added +43.4 ppm. and water was removed by means of a Dean-Stark trap (about Anal. Calcd for C,oHss02P: C, 83.31; H, 5.77; P, 5-37. '/a of the solveiit was removed), and the temperature reduced Found: C, 83.29; H, 5.66; P, 5.64. t o 50', a t which time 2-bromopropene, 9.0 g (0.075 mol), was Additional material recovered from washings of the initial introduced. The mixture was allowed to stir 24 hr at 50' and residue was shown to be 22a, 2.8 g (21%), mp 229-230' (described elsewhere), identified by comparison to authentic sample. then 72 hr at 200' (mantle temperature) and filtered to remove lumps of black intractible material. Benzonitrile wm removed Salts of 3- and 4-Substituted Phospho1anes.-The salts were by steam distillation and the salt extracted with chloroform. made by dissolving or suspending the phospholanes in chloroConcentration followed by several precipitations in ether and form, adding excess 6 M HBr, and shaking vigorously for about several recrystallizations from ethyl acetate-ether gave 10, 1 hr. The chloroform solution was then separated, washed with water, dried (MgSOd), concentrated, dropped into a 1 : l 13.4 g (70%), mp 195-197' (lit.14 196.5-197.5'). This is a modification of a previously reported method which failed to ether-hexane mixture, filtered, and recrystallized from a chlorogive the desired product .14 form-hexane mixture. Products and yields are as follows. 5-Benzoyl-3-methyl-2,2,2,5-tetraphenyloxa-2-phospholanes Phospholane 7a gave pure salt 19a: 9870; mp 246-248'; ( l l a and llb).-To a prereacted mixture of NaH, 0.65 g (0.015 ir (CHCla) v 1110 (s, CP), 1230 (m), 1680 ( 6 , C-O), 3150 cm-l (s, OH); nmr (CDCla) 6 1.0 (d, 3, CHa), 2.3-3.2 (m, 2, HCCHP), mol), and 50 ml of dry DMSO was added 1, 3.2 g (0.015 mol), and then (slowly) 10, 5.8 g (0.015 mol), as a solution in 50 ml of 4.4-5.1 (m, 1 , CHP), 6.2 (s, 1, OH, disappeared on shaking DMSO. A light green precipitate formed slowly and was colsample with DzO), 7.2-8.1 ppm (m, 25, CeHs). Anal. Calcd for C36H&PBr: C, 70.59; H, 3.42; Br, 13.42. lected by filtration after 4 hr. The residue was washed with Found: C, 70.35; H, 5.31; Br, 13.65. DMSO, water, and ether. Fractional crystallization of this product from a chloroform-hexane mixture afforded two similar Phospholane 7b gave pure salt 19b: 90%; mp 216-219'; jr (CHCI,) Y 1110 (9, CP), 1235 (m), 1670 (s, C=O), 3100 cm-' oxaphospholanes, l l a and l l b , the bulk of the latter being recovered by concentration of the clear solution after addition of (s, OH); nmr (CDC13) 6 0.6 (d, 3, CHI), 2.6-3.2 (m, 1 , CH), hexane failed to bring about further precipitation. 3.3-4.0 (m, l ) , and 4.0-4.8 (m, 1, CHnP), 5.3 (s, 1, OH, disappeared on shaking sample with DzO), 7.0-8.1 pprn (m, 25, Although the melting points and mixture melting point of these compounds are identical, 179-181', the nmr spectra show that C6Hd Anal. Calcd for C~,,FTa~OgPBr: C, 70.59; H, 5.42; Br, they are different. Yields were 2.6 g of l l b (34%) and 2.8 g 13.42. Found: C, 70.70; H, 5.45; Br, 13.80. of l l a (36%). Integration of the nmr spectrum of the original Phospholsne 9 gave pure salt 20: 88%; mp 178-182'; ir product mixture (methyl peaks) gave relative percentages of 48 (CHCla) Y 1110 (s, CP), 1230 (s), 1670 (s, C=O), 3100 ern-'' ( l l b ) and 52 ( l l a ) . 5-Benzoyl-3-methyl-2,2,2,5-tetraphenyloxa-2-phospholanes (s, OH); nmr (CDCla) 6 2.0 (s, 1, OH, disappeared on shaking with D20),3.4-4.9 (m, 3-HCCI-lzP), 6.6-8.2 ppm (m, 30, CGHS). (Ilb): ir (CHCla)v 1066 (s, CO), 1100 (s, CP), 1200 (m), 1230 Anal. Calcd for C,oHarOzPBr: C, 73.06; H, 5.21; Br, (m), 1250 (m), 1650 cm-1 (s, C=O); nmr (CDCla) 6 1.2 (d d, 12.15. Found: C, 72.83; H, 5.28; Br, 12.46. 3, CHa), 1.6-2.5 (m, 1, HB, CCP), 2.9-3.8 (m, 2, HA, CCHP), Phospholane l l a gave pure salt 21a: 97%; mp 240-244'; 6.9-8.0 ppm (m, 25, CeHs). ir (CHCl,) v 1118 (7, CP), 1240 (s), 1680 (s, C=O), 3150 cm-' When the above sample was shaken with DzO and allowed to stand 10 min, the following spectrum was observed: 6 1.2 (d, (s, OH); nmr (CDCl3) 6 0.8 (d d, 3, CHI, J H ~ C C P = 21 HE), 1.9-3.5 (m, 3, CHn plus OH, broad singlet disappeared and 3, CHa, JH~CCP = 21 Hz), 2.0 (d d, 1 , H B CCP, ~ J H B P P = 27, integration dropped to 2 when sample was shaken with DsO), J H A ~ H 13 B ~Hz), 3.3 (d d , 1 HA~CCP, JHA'P = 20 Hz), 6.9-8.0 5.0-5.7 (m, 1, CHP), 6.9-8.2 ppm (m, 25, CGHS). ppm (m, 25, C6Hs). This spectrum was unchanged after 24 hr. IIP nmr showed 6 +48.5 ppm. Anal. Calcd for CasHI20zPBr: C, 70.59; H, 5.43; Br, 13.42. Found: C, 70.42; H, 5.30; Br, 13.81. (13) The authors wish t o thank R . Cambell for refinement of this proPhospholane l l b gave salt 21b, 92%, containing, however, cedure. I
(14) J. G . Thompson, Doctoral Dissertation, University of Delaware, 1968.
(15) E. E. Schweiaer and A. T. Wehman, J . Chem. Soc., 3.13 (1971).
REACTIONS OF PHOSPHORUS COMPOUNDS
J. Org. Chem., Vol. 36'1 No. 1B118?'i 2249
2,3-Diphenyl-5-methyl-2,5-dihydrofuran (14).-To a suspension of sodium hydride (2.1 g, 0.05 mol) in 50 ml of anhydrous ether was added benzoin (l), 12.8 g (0.05 mol), and then a solution of salt 6, 19.2 g (0.05 mol), in 150 ml of dry CH,CN. After a 48-hr reflux the mixture was filtered and the residue was washed with water and ether and air-dried to give 8.3 g of 7a, plus a small amount of 7b (32% total). Dilution of the filtrate with 500 ml of water followed by extraction with two 100-ml portions of ether, drying (MgS04), concentration, and dropping into 400 ml of hexane gave a gummy precipitate and a clear yellow solution which was decanted. The gum solidified on stirring and proved to be slightly impure triphenylphosphine oxide (4), 7.8 g (56%). Concentration of the decantate and chromatography on silica gel afforded 3.6 g of a light yellow gum which proved to be pure 14 (30%), identified by its ir and nmr spectra. However, after repeated at tempts to crystallize this product, only 1,2-diphenyl-2-pentene1,4-dione was recovered. Recrystallization from cold C C 4 hexane gave colorless needles, mp 119-121', 2.6 g (68%). 2,3-Diphenyl-S-methyl-2,5-dihydrofuran (14): ir (CHCls) Y 1070 om-l (s, COC); nmr (CDClt) 6 1.2 (d d, 3, CH3), 4.85.2 (m, 1, CH), 6.2-5.8 (m, 2, ArCH C=CH), 7.6-6.9 ppm (m, 10, CeH6). Anal. Calcd for CI~HIBO:C, 86.40; H , 6.83. Found: C, 86.74; H, 6.58. 1,2-Diphenyl-2-pentene-l,4-dione:ir (CHCla) Y 1680 (s, C=O), 1580 (s, C=C), 1280 (m), 1190 (s), 985 cm-l (m); nmr (CDCla) 6 2.2 (8, 3, CHs), 6.8 (s, 1, C=CH), 7.2-7.7 and 7.8-8.0 ppm (m, 8, and m, 2, respectivcly, CsH5). Anal. Calcd for C&40z: C, 81.58; H, 5.64. Found: C, 81.36; H, 5.71. This reaction was repeated using salts 8, 10, and 12. (A) Reaction of 1 and salt 8 gave compound 9 (8%), 2,3,5triphenylfuran [4%, mp 93- 94' (lit.IB 92-93')], 1,2,4-triphenyl2-butene-1,4-dione [6%, mp 129-129.5' (1it.l' 129')], and 2,3,5-triphenyl-2,J-dihydrof~ran~~ (15) 150%; ir (CHCL) Y 1030 (m), 1060 (8, COC), 1070 (m), 1600 cm-l (m, C=C); mp 78-80' (lit.'' 79-40'); nmr (CDCl,) S 5.8-6.1 (m, 1, C= CH), 6.2-6.4 (m, 2, HCOCH), 6.9-7.16 ppm (m, 15, CoH5)]. Anal. Calcd for CzzHlsO (15): C, 88.59; H, 6.00. Found: C, 88.40; H, 5.90. (€3) Reaction of 1 and salt 10 gave compound lla containing 36% of llb (34% combined) and 2,3-diphenyl-4-methy1-2,5dihydrofuran (16) [28%; ir (CHCla) Y 1070 cm-l (s, COG); TABLE I11 mp 72-74'; nmr (CDCla) 8 1.7 (broad doublet, 3, CHI, J = CONVERSION OF OXAPHOSPHOLANE SALTSTO OXAPHOSPHOLANES 2 Ha), 4.7 (m, 2, CHZ), 6.0 (m, 1, CH), 6.9-7.7 ppm (m, 10, Yield, C&)l. Product % Salt Anal. Calcd for CI,Hl6O (16): C , 86.40; H,6.83. Found: 7a 95 19a C, 86.44; H, 6.79. 7b 85 19b (C) Reaction of 1 and salt 12 gave compound 13 (76%). There was no discernible evidence for the formation of the di9 90 20 hydrofuran 17 or any products resulting directly from it. lla 88 21a 21b 21a (21%) llb lla (20%) 85 Registry No.-7a, 30697-85-9; 7b, 30697-86-0; 13 67 22a 22a 22b (50%) 13 74 9, 30745-05-2; lla, 30697-87-1; llb, 30697-88-2;
21% of salt 21a, mp (of the mixture) 138-145'. These salts were inseparable by column chromatography, tlc, and fractional crystallization and showed no tendency to isomerize to one pure salt under the influence of base. For an 80% pure sample of 21b, ir (CHClr) was identical with that described for = 20 Ha). 21a; nmr (CDCla) 6 1.5 ppm (d d , CHS, JH~CCP Other absorptions cannot be identified with certainty due to contamination with 21a. Anal. Calcd for C~6H3202PBr(mixture of isomers): C, 70.59; H, 5.42; Br, 13.42. Found: C, 70.27; H, 5.45; Br, 13.63. Phospholane 13 gave a 1:1 mixture of salts 22a and 22b 96%. The ratio was determined by nmr integration. Two spots were shown by tlc. When separation of these components was attempted with refluxing benzene in a Soxlet extractor, pure salt 22a was recovered quantitatively, mp 234-236'. Heating the mixture a t 100' in DMSO containing a trace of potassium tertbutylate also afforded pure 22a after 48 hr. For 22a: ir (CHCl) Y 1110 (8, CP), 1230 (s), 1670 (s, C=O), 3150 cm-l (9, OH); nmr (CDCla) 6 2.1 (broad s, 1, OH, disappeared when sample was shaken with DzO), 2.9-3.5 (m, 3, HtCCHP), 6.8 (broad s, 5, CaHb), 7.0-8.0 ppm (m, 25, C ~ H S ) . Anal. Calcd for C40HarOzPBr: C, 73.06; II, 5.21; Br, 12.15. Found: C, 72.83; H , 5.28; Br, 12.46. A 90% pure sample of salt 22b was obtained by fractional crystallization from chloroform solution in an ether atmosphere, mp 191-193'. For 22b: ir (CHCla) was essentially the same as that described; nmr (CDCls) 6 2.5 (broad s , 1, OH, disappeared on shaking with DzO), 2.8-3.5 (m, l ) , and 3.6-4.2 (m, 1, CH$), 5.2-5.8 (m, 1, CHP, appears as broad d d), 6.8-8.0 ppm (m, 33, C6H5, including a small peak at 6.8 corresponding to that in the spctrum of 22a). Anal. Calcd for C40Hs40zPBr (1:l mixture of 22a and 22b): C, 73.06; H, 5.21; Br, 12.15. Found: C, 72.99; H, 5.04; Br, 12.50. Conversion of Salts to Phospho1anes.-The salts were dissolved in a small amount of methanol and 10 M NaOH was added dropwise until the solution became turbid. On standing the phospholanes crystallized and water was added at intervals until no further precipitation cccurred. The crystals were recovered by filtration and washed with water and ether. Results are in Table 111.
+ +
+
+
Pyrolysis of 3- and 4-Substituted Oxaphospholanes A.Samples of oxaphospholanes 7a, 7b, 9, lla, llb, and 13 were injected into a vpc (column W-98 on Chromosorb, 200'). The only peaks recorded corresponded to benzil (23), triphenyl phosphine (24), and, in the cases of 9 and 13, styrene. These products were collected and their identities confirmed by comparison to authentic samples. B.-Samples of each of the above phospholanes (except 13) were recovered unchanged from a 4-day reflux in krt-butyl alcohol, although very small amounts of benzil, triphenylphosphine, and, triphenylphosphine oxide could also be detected. Compund 13 was recovered in only 32% yield and significant amounts of benzil (23, 30%), triphenylphosphine (24, 17%), and triphenylphosphine oxide (4, 18%) were isolated. Styrene was detected by vpc but no other products were observed.
13, 30697-89-3; 14, 30697-90-6; 15, 30697-91-7; 16, 30697-92-8; 18, 30697-93-9; 19a, 30697-94-0; 19b, 30697-95-1; 20, 30697-96-2; 21a, 30697-97-3; 2 lb, 30697-98-4; 22a, 30697-99-5; 22b, 30698-00-1; 1,2-diphenyl-2-pentene-l14-dione, 24105-46-2. Acknowledgment. -We gratefully acknowledge support by a Public Health Service Grant (CA11000-06) from the National Institutes of Health. (16) F. R. Japp and F. Klingemann, C h e n . Be?., 21, 2932 (1888). (17) F.R.Japp and N. H. J. Miller, J . C h e n . SOC.Trans., 35 (1885). (18) L. F. Chelpanove. and V. A. Kormer, Zh. Obshch. K h i m . , 29, 2348 (1959). However, the spectral data reported for this compound leaves some doubt as to the validity of the previous characterization.
